Semantic aware online detection of resource anomalies on the cloud

As cloud based platforms become more popular, it\ud becomes an essential task for the cloud administrator to efficiently\ud manage the costly hardware resources in the cloud environment.\ud Prompt action should be taken whenever hardware resources are\ud faulty, or configured and utilized in a way that causes application\ud performance degradation, hence poor quality of service. In this\ud paper, we propose a semantic aware technique based on neural\ud network learning and pattern recognition in order to provide\ud automated, real-time support for resource anomaly detection.\ud We incorporate application semantics to narrow down the scope\ud of the learning and detection phase, thus enabling our machine\ud learning technique to work at a very low overhead when executed\ud online. As our method runs “life-long” on monitored resource\ud usage on the cloud, in case of wrong prediction, we can leverage\ud administrator feedback to improve prediction on future runs.\ud This feedback directed scheme with the attached context helps\ud us to achieve an anomaly detection accuracy of as high as\ud 98.3% in our experimental evaluation, and can be easily used\ud in conjunction with other anomaly detection techniques for the\ud cloud